benchmarks.site

Public leaderboard · Aggregated leaderboards

LLM Stats

A website that gathers public test results, prices and speeds for hundreds of AI models and ranks them with a single combined score.

Leaderboard

Top 3 right now

  1. 1stClaude Fable 5.260.8
  2. 2ndClaude Mythos 5.259.3
  3. 3rdClaude Opus 5.558.5

What it measures

Collects benchmark scores, prices and context sizes for many models into comparable tables.

How to read the numbers

LLM Stats composite score (higher is better) from public benchmark results — higher is better.

Results

15 rows · higher is better (LLM Stats score)CSV
1Claude Fable 5.2Anthropic60.860.049.846.3
2Claude Mythos 5.2Anthropic59.353.444.544.2
3Claude Opus 5.5Anthropic58.557.447.843.6
4Gemini 4 CarbonGoogle57.454.045.743.7
5Claude Sonnet 5.5Anthropic56.755.446.041.9
6GPT-6 AstraOpenAI56.355.043.942.3
7Kimi K3.1 ProMoonshot AI54.453.340.539.9
8Claude Mythos PreviewAnthropic53.654.642.535.1
9Gemini 4 ArgonGoogle53.351.241.339.6
10Muse Wonder FullMeta52.644.842.040.4
11Claude Opus 5Anthropic52.552.537.336.7
12GPT-6.1 SolOpenAI52.249.637.635.8
13Claude Fable 5.1Anthropic52.049.834.437.1
14Muse Wonder FullMeta52.040.140.940.0
15Claude Fable 5Anthropic51.949.841.637.0

Note: Homepage leaderboard table (page says 'updated continuously'); includes some unreleased/preview models. Scores are an aggregate of other benchmarks' public results.

Source: LLM Stats (llm-stats.com) · llm-stats.com · retrieved 2026-11-17

Markdown · More Aggregated leaderboards